A higher setting for the animation slider seems to allow the engine pick greater distances in which to create the animation/engagement. For example, on 100, the blocking would likely have very noticeable suctioning. 100 Tackling has made ball carriers slow down so the defender can be drawn to him to begin the tackling animation.
For the ratings sliders, it's likely a boost/drop in the rating either directly or somewhere deeper in the engine when it determines how things develop as the play goes on.
The thing about 50s, at least for me, is that you're counting on the engine to be realistic (both in stats/results, and in challenge/gameplay) at 50s. If the engine has created "too much" success or failure by default, then 50s will express that. How much is too much/too little depends on a lot of factors - not the least of which is the human at the controller's playstyle/abilty/"stick skills"/etc
Not to mention you have to think of the interactions. For example, with Break Tackle at 50 and Tackling at 0, you're getting a lot of broken tackles with tackles engaging basically if the players run into each other. I would lower Break Tackles even if 50 was the most realistic. That high a difference might also create a lot of "pro-tak moments" where the pile is broken for a TD.
I've read that they should be within 5 of each other to minimize it and since I don't like 55 tackling, I lowered break tackles a ton and let the difficulty in creating solid engagement (tackling in this case) make more chances for breaking tackles, so to speak. Not sure if the game counts them as broken tackles, but gameplay-wise, they look like broken tackles.
I guess you could say my personal "rule" is: If I lower the amount of engagement, I must also lower the failure rate of that engagement. So for my sliders, if I lower the amount of tackling engagement, I lower the success rate/impact of breaking tackles. I don't want it to be two-hand touch out there where a high TRU/ELU runner can't shed tackles or where weak tackling/position doesn't matter, but don't want gang tackles when they do occur to do more harm than good. Instead, I want ball carriers to be able to fall forward unless taking huge direct hits, or faced with higher-end tackle ratings.
I break that rules with run blocking and block shedding, though maybe it's just my hatred of how run blocking works in this game that's making me biased. Every time I try a run blocking of or near 0, I think run plays turn into a mess (or more of a mess depending on how you look at it), but again, probably just me. Considering running can be stopped, sometimes easily, at 0 block shedding, it can work. Depends, perhaps, on if you want penetration to be the #1 factor in stopping the run (higher block shedding) or gap control (lower block shedding - have to create free runners and only the high BSH players get penetration).
Pass blocking could work (it seems to run on a different dynamic than run blocking), but since players can reach out with their arms and such, I give a little leeway. Even with run blocking, I'd give a little. Maybe try something like 5 or 10 to begin with for the three "suction sliders" (since guys can reach with their arms to grab and make tackles too)
Interceptions seems to be touchy. Too high and you'll have a pick fest. Too low and you'll see lots of dropped INTs, even on blantatly bad passes. The opportunities also vary based on pass reaction (players get in/stay in better position to make plays on the ball). Not to mention LBs seem better than DBs at getting picks (could be an AWR thing, or maybe passes thrown in the area of LBs are not as hard to catch so they can make the catches more often. Velocity and trajectory of the pass matters.
I would keep it at or under 20 to begin with, and then see what you like and don't like. That said, if you play on slow with a lower QB Acc, you might want to go down to around 10. Slow gamespeed slows down the ball in the air too, and that seems to impact things as well!
WR catching, I would put around 50. I wouldn't want WR to get even more passive about going after the ball than they already are, but wouldn't want them to turn into supermen than can catch horribly thrown balls in a single warp and hold on to it constantly when pounded when they warp right into the crosshairs of a safety.
QB Accuracy is another touchy one because the human player has the most impact here on offense. He/she is always the QB after the snap (barring trick plays) so the human's abilities impact things as much as anything else. For the CPU, I would lower it below 50 if only to keep the top QBs in line, especially with completion percentage. The highest Comp% this season so far is about 70% (Brees, Manning, etc). I wouldn't want the top QBs (much less those below them) to regularly get above that number.